Scribd
Upload
8 views

Assignment1

Uploaded by

Mr Unknown
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
8 views

Assignment1

Uploaded by

Mr Unknown
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

12/16/24, 10:57 PM Assignment

Assignment
Deadline for submission - 15 Dec 2024, 11:59:59 PM

1. Create a Google Colab Notebook


Your Colab notebook should include the following:
1. Title and Introduction

• Add a descriptive title (e.g., "Image Generation and Background Replacement


using Stable Diffusion").

• Write bullet points to include the project goals and what the notebook
covers.
2. Comments and Explanations
Add detailed comments ( # ) throughout the code to explain:

• The purpose of each section.

• The methods and functions you use.

• Any parameters or configurations you set.

2. Model Selection
Explore different image generation models and their functionalities. Focus on
models that support:

• Text-to-Image

• Image-to-Image

• Inpainting
Specifically, look into the following versions of Stable Diffusion:
1. Stable Diffusion 3.5 [You can use flux as well]
2. Stable Diffusion 3
Discuss the strengths and differences between these models, and document your
findings in word doc.

3. Adapter Fusion
Include adapters (face/style) that can be fused with Stable Diffusion models to
enhance their capabilities.

https://festive-spectrum-d2d.notion.site/Assignment-159433b3864080d491a0e3d511d353e2 1/3
12/16/24, 10:57 PM Assignment

4. Quantization
Select either Stable Diffusion 3 or Stable Diffusion 3.5 and quantize the model to
nf4 format.

• Show the steps to quantize the model. [Please skip this and use quantized
weights using diffusers]

5. Object Replacement Feature


Develop a feature to replace the background of an image using the quantized
model.

• Ensure the mask for object/human replacement is auto-generated (no manual


masking) via a prompt.

• Write code that demonstrates the entire process:


1. Load the quantized model.
2. Automatically generate a mask.
3. Replace the object with a new one.
4. Display the results.

7. Deliverables
1. Colab Notebook Link: Share a link to your Google Colab notebook.

After completing the assignment please fill this form to submit your assignment
→ https://airtable.com/appCoFESellVg29ax/pagUpZ8yrPTBuDEmc/form

Deadline for submission - 15 Dec 2024, 11:59:59 PM

8. Additional Tips
• Research Thoroughly: Spend time understanding the models, quantization
techniques, and background replacement methods.

• Code Quality: Ensure your code is clean, well-commented, and follows best
practices.

• Test and Validate: Verify that all functionalities work as expected before
submission.

https://festive-spectrum-d2d.notion.site/Assignment-159433b3864080d491a0e3d511d353e2 2/3
12/16/24, 10:57 PM Assignment

• Visual Results: Include before-and-after images to demonstrate the background


replacement feature.

• For any clarification please reach out to [email protected]

https://festive-spectrum-d2d.notion.site/Assignment-159433b3864080d491a0e3d511d353e2 3/3

You might also like